The text of the Quran is principally dependent on the Arabic language. Therefore, improving the security and reliability of the Quran’s text when it is exchanged via internet networks has become one of the most difficult challenges that researchers face today. Consequently, the diacritical marks in the Holy Quran which represent Arabic vowels (
Nowadays, the tamper detection of text documents has increased significantly for secure data transmission over the internet [
Capacity, imperceptibility, robustness and safety are the watermarking criteria. Capacity is the total number of hidden bits while an imperceptibility location is used to measure the difference between the original location and the watermarked location, noting any additions to the original location. Robustness is the capacity to remove or detect a watermark after attacking the watermarked location. Finally, the security requirement is the challenge of removing a watermark without destroying it [
Text watermarking is one of the media’s criteria of digital watermarking and has become crucial due to developments in information concealment. It refers to the process of embedding the watermark inside a text document as a way of protecting in terms of the authentication of the document. The protection of such information is necessary to preserve the watermarked text so that it is the same as the original text without change. In the domain of text watermarking, there are three major issues which must be resolved: imperceptibility, capacity, and security [
The Quran was written in Arabic, so in order to conceal information within it, one needs to know the features of the language. Text watermarking has many duties, some of which include the use of the “kashida” letter extension (one of the features of the Arabic language) and the Arabic vowels
In this paper, the author presents a scheme for improving the embedding imperceptibility and security through the use of tamper detection and location authentication of a Quranic text watermarking scheme based on vowels with kashida using the exclusive-or (XOR) and queuing techniques. The proposed scheme utilizes the Queuing technique which was chosen in order to check the embed phase for the tamperage of words forms the text of the Quran’s text. As a result, when any of type of attacks such as deletion, insertion, or reordering is detected one can then truly say whether or not the text was, indeed, tampered with, check by extract phase. The main contribution is the development of an improved tamper detection scheme. Thus, this contribution is different from other similar approaches of existing literature due to the addition of the location of authentication in Quranic text watermarking. The proposed scheme was implemented to detect tamperage by a hacker that tried to manipulate the watermarked Quranic text. The use of the hybrid technique refers to using both the XOR and the Queuing techniques which make the proposed scheme more secure in terms of fast and accurate tamper detection. The proposed scheme was able to detect the location of the insertion, deletion and reordering of the tamperage accurately.
This paper contains five more sections. Section 2 provides related work. Section 3 presents the proposed methods and evaluation. Section 4 discusses the comparison and results, and finally, Section 5 offers the researcher’s conclusions.
Obtaining the authentication and credibility of users and information through the internet is vital. However, ensuring the Quran’s text remains authentic is even more vital. The authenticity and originality of verses in digital media can be challenging to preserve especially as it is prerequisite for accomplishing this feat in a quick manner. The proposed method for implementing a Quranic authentication system is based on information retrieval techniques and hashing algorithms. The evaluation of different characteristics may impact the accuracy of the authentication process. Results showed that hashing verification is a good candidate for automatic authentication with high confidence. As a result, rather than focusing on encryption, the primary focus is on authenticating information. However, finding proper Quranic documents online for this research is difficult; consequently, the relationship between information providers and readers is not typical e-commerce, one-to-one relationship in which typical hash algorithms can be sent to the known receiver. Nevertheless, private keys intended for a receiver will not be applicable in our scenario where the tool is expected to quickly check the Internet for authentic Quranic documents that have not been tampered with [
Tamper detection in plain text papers uses a zero-watermarking algorithm. The algorithm produces a watermark dependent on the quality of the text which can later be obtained using the extraction algorithm to define the tampering position in the document. Watermark matching patterns and watermark distortion rates on several text samples of varying duration are used as measurement parameters. Findings show that this algorithm still detects tamperage even when the amount of it is limited. With the limited volume of attack, the presence of a watermark and the reliability of the details are, therefore, difficult to determine [
A zero-watermarking method used for authentication utilizes the characteristics of the Arabic characters without changing the file. The watermark key focuses on the characteristics of the original verses For each verse of the Quran in the initial stage, a key is produced which checks the name and number of the surah along with its
Some researchers suggest using a tamper locating algorithm for authentication of the content found in .docx documents. The authentication information that is not connected to the text content is stored in the key setting file called document.xml through the segmentation of the display characters. Identifying the quality of the text content can detect whether or not the embedded watermark is similar to the authentication watermark. Tests show that the algorithm is very sensitive to any change and can very well identify the manipulated places [
Authentication of image content is primarily used to determine the validity of image content and can easily locate the tampered area in order to determine the motive for actions. Therefore, for content-preserving processes such as noise, compression, and replacement methods, the authentication algorithm must be robust and made vulnerable to malicious attacks. Furthermore, semi-fragile digital watermarking techniques can, to some degree, differentiate between content-preserving activities and malicious tampering attacks [
Referring to multiple watermarking systems to secure ownership and tamper detection, this scheme robustly embeds the watermark details used for authenticating copyright by the synonymous replacement process; thus, it uses a double fragile watermark to achieve the detection and position of manipulations. The suggested system has a strong benefit when it comes to its implementation; however, are already some shortcomings such as the watermark’s robustness, the precision of finding the tamperage when dealing with a more complicated attack, both of which need further developments [
According to the novel CNN-based security-guaranteed image watermarking generation scenario for smart city applications, the content-based watermark synchronization scheme recognizes watermark embedding positions via stable image feature points in the present anti-geometric attack watermarking algorithm, embeds the watermark in a local neighbourhood feature point, and uses the feature points to locate the watermark. These methods demonstrate increased robustness. The gradient path distribution of the neighbouring pixels of a feature point is used to determine the direction information for each feature point. In practice, in a neighbourhood window based on the feature points, samples are generated, and a histogram is used to measure the gradient directions of the adjacent pixels. Thus, with the main direction of the feature point, the peak of the histogram represents the primary direction of the neighbourhood gradient. The model achieves enhanced robustness with the incorporation of CNN [
The quaternion Fresnel transform (QFST), computer-generated hologram (CGH) and two-dimensional Logistic-adjusted-Sine map (2D-LASM) techniques, a novel four-image encryption scheme is proposed. Four images are in the proposed process, Holistically processed in a vector manner by first using QFST. Then, using virtual RPM, phase-shifting interferometry (PSI) and Burch’s coding process, the encrypted CGH of the input complex amplitude, which is constructed by the components of the QFST-transformed images, is generated [
The RPMs are in order to avoid using the entire RPMs as the decryption keys 2D–LASM produced. LASM’s initial values are used as decryption keys. CGHs have the advantage of digital fabrication, storage and transmission compared to traditional optical holography. Subsequently, the obtained CGH is permitted based on LASM to improve security. Numerical simulations show the feasibility and performance of the proposed system. The encrypted hologram is scrambled based on 2D-LASM to enhance the security of this paper and weaken the correlation. The validity of the proposed technique of image encryption is demonstrated by experiments [
The 2D Fuzzy C-Means (FCM) algorithm has been widely used for segmenting medical images. Throughout the years, various extensions of it were proposed. A modified version of FCM was therefore proposed for segmenting 3D medical volumes, which has rarely been implemented for 3D medical image segmentation. We present a parallel implementation using the Graphics Processing Unit of the proposed algorithm (GPU). Researchers state that when working with 3D models, reliability is one of the primary problems when using FCM for medical imaging [
A hybrid parallel implementation of FCM is therefore suggested for removing volume objects from medical files. Using actual medical data and virtual phantom data, the proposed algorithm has been validated. The key factors for the device validation were segmentation accuracy of predefined datasets and actual patient datasets. To demonstrate the efficiency of each implementation, the processing times of both the sequential and the parallel implementations are calculated. The results obtained conclude that the parallel implementation is 5x faster than the sequential version of the same operation. Concentrate on automatically segmenting oblique slices from 3D volumetric data. To minimise the processing time here as much as possible, an acceleration mechanism using the GPU will be considered [
Watermarking systems have been implemented before, but authentication and tamper detection are still a major research issue. A double image-based watermarking scheme was introduced using the Local Binary Pattern (LBP) to protect multimedia documents from unauthorized alteration. The suggested approach involves the following procedure: the host image is partitioned into non-overlapping blocks during watermark encoding. This is followed by the generation of the process vector (S) using LBP and secret watermark bits to execute XOR. A two-bit authentication code is created from the (S) vector and embedded in a dual image based on a shared secret key (ÿ). The findings are compared to state-of-the-art approaches to show the usefulness of our suggested system. It is noted that the proposed scheme is secure and robust against different standard attacks whereas it can detect the integrity of messages inside the watermarked item [
Different scenarios have recently been proposed by the researchers depending on the insertion of variable numbers of kashida per word. Such strategies have produced better results in terms of capacity and security than the previous methods; however, they also reveal a noticeable weakness in the process of retyping [
Previous literature shows research has been conducted on text watermarking approaches and methods that were proposed for several purposes of information security. Some research proposes a robust content authentication method for paper text documents, solving the problem of paper document content integrity verification, tamper identification and tamper position locating. Furthermore, in this method, the watermark information does not depend on the additional carrier, and the text document is embedded with the watermark visible. The watermark is embedded during the process of a printout, which avoids the risk of malicious tampering of the watermark information. As a result, this method is seen to possess high-security performance. The watermarking algorithm is robust against the print-and-scan attack and has a high capacity with low accuracy. In addition, the minimum string edit distance algorithm is used to compare the difference between the authentication watermark signal sequence extracted from the scanned image and the one calculated in real-time, through which it determines whether or not the paper document’s text content has been changed, and if so, it accurately locates the place where it was tampered with [
The Intrusion Detection System (IDS) is an application that detects malicious attacks or data breaches within a network. As a vital network security element, IDS has been frequently used in recent years. This analysis aims to define the best-fit solution, that would reduce the number of features substantially. Furthermore, with less processing time, the method would lead to high classification accuracy. We propose a hybrid feature selection model that combines the strengths of the legacy filtered and wrapper selection mechanism to reach this aim. This proposed amalgam of approaches is supposed to efficiently pick the optimal set of intrusion detection features. Using correlation feature selection (CFS) along with three different search techniques known as best-first, greedy stepwise and genetic algorithm, the suggested hybrid model was carried out [
In order to evaluate each of the features that were first selected by the philtre process, the wrapper-based subset evaluation uses a random forest (RF) classifier. On both the KDD99 and DARPA 1999 datasets, the reduced feature range was checked in a supervised setting using an RF algorithm with ten-fold cross-validation. The outcome shows that, in terms of detection time, accuracy, and detection rate, choices made by the hybrid feature had better performance improvements. A low false alarm rate was reported, as well [
The suggested zero watermarking utilized the material features of the text to create a watermark instead of embedding the watermark in the text using the structural variable and word length, that is common to all kinds of texts. It consisted of two stages: text embedding and extracting. The watermark was generated with the data holders, and the extraction was done with the certifying authority (CA) [
A technique was developed that used a text document as an input for manipulation detection, and the watermark was created based on the HSW (hybrid structural component and word length) approach. A watermark extracted pattern was registered with the (CA) where the attacker could change the document’s contents. During the tampering identification, the extraction method was used to retrieve the watermark and the template fitted the recorded (CA) model. A decision on the interference was made that depended on the degree of pattern matching with MD5 compression. The evaluated attacks including the deletion, insertion, and reordering using this method achieved high accuracy than other works and. However, this approach could not update the detailed ownership of the source in the generated watermark information [
The proposed zero-watermarking technique known as watermark arrangement is based on the Markov Model Level 4 Word Mechanism (ZWAFWMMM) and is used for authenticating information and detecting tamperage within the Arabic text. It is an effective model as ZWAFWMMM adopts a hybrid system. Nevertheless, due to the complicated nature and structure of the Arabic language, the basic curriculum uses conventional techniques which lack the capacity to provide effective solutions to the Arabic text. The findings of the experiment reveal that ZWAFWMMM is more sensitive to all forms of tampering and highly accurate when it comes to the low capacity of tamper detection [
As mentioned before, the Holy Quran is a religious book that plays an extremely vital role in the life of Muslims as the main decisions of Islamic jurisprudence and, indeed, the everyday life of Muslims are based on the instructions written in the verses of the Holy Quran. Ordinary Muslims cannot judge the authentication of the verses of the Holy Quran if the verses have been tampered with. In fact, it requires intense attention, extensive knowledge, and dedicated efforts to be able to tell the difference between true Quranic verses and ones that have been tampered with, especially when one or more words are missing from the recitations. Typically, the authenticity of an online quote attributed to the Quran can be confirmed by making a comparison between the online quotations and the original text of the Quran [
References | Domain | Methods | Type of Attacks | PSNR (dB) | NCC/% | Remarks | |
---|---|---|---|---|---|---|---|
[ |
Text document image | The Pascal Triangle | Text Modification Attack | 54.95 | 0.98 | Increased the imperceptibility and security with less capacity | |
[ |
Quran text | XOR diacritics of the special characters | Message Modification Attack | 56.34 | NA | Increased the capacity and solve the complexity | |
[ |
Quran text | The moon and sun letters | Message Modification Attack | 61.16 | NA | Increased capacity with low security | |
[ |
Text document | LSB algorithm | Text Modification Attack | 62.46 | NA | Increased imperceptibility with less security | |
[ |
Arabic text | Inserts the pseudo-space and other three small or zero-width spaces | Text Modification Attack | NA | NA | High imperceptibility and capacity with low security | |
[ |
Arabic text watermark | Adding diacritics into any Arabic letter | Text Modification Attack | NA | NA | High capacity with less imperceptibility and security | |
[ |
Arabic text Hadiths | Counting based secret sharing with kashida | Text Modification Attack | 52.05 | NA | Increased the imperceptibility and security with less capacity | |
[ |
Arabic Text documents | Kashida with diacritics | Text Modification Attack | 59.04 | NA | Considered the authentication with a high degree of capacity |
The methodology used in this research is divided into four phases: The first phase is the pre-processing phase which, as its name suggests, is responsible for the preparation of the hosting media such as the Quranic text in this case, and a secret bit (data that was hidden in the text). The second phase includes the embedding of the secret bit within the Quranic text. The third stage involves the extraction of the data including the attack process and the final phase is the performance evaluation of the scheme through the use of various measures. The details of each phase are discussed below.
In this phase, the watermark is converted from the binary image into the sequence of bits by decoding it with certain conditions (e.g., inserting a kashida in the next letter after the vowel). Next, the binary image is scanned to sequentially find the byte. If the byte value is 255, then the secret bit is 0. After finishing the binary image scan and byte assignment, vector 1 and 0 are produced. During this phase, the original Quranic text is called the cover text. This text consists of various characters that are given a unique ASCII code with a decimal value. This decimal’s value is useful for determining the condition of embedding which is known in advance. Each location is stored in the vector-based condition accordingly and produces the vector for the position as shown in
The proposed scheme is based on the vowels to be embedded in the Arabic text where the three vowels are used. These letters are selected because they are the most redundant characters in the whole text of the Quran. The embedding process is comprised of the preparation phase for both watermark and covers of the Quranic text. For watermarked text, all the bytes are converted into bits and then stored in the vector. The cover of the Quranic text should scan first for counting several vowels and check the condition (inserting a kashida in the next letter after the vowel) that is appropriate with the embedding protocol.
The embedding strategy is accomplished in the following six steps:
Prepare both watermark and cover Quranic text. The preparation step is necessary for any watermarking scheme which is comprised of two elements. Prepare the watermark for conversion into binary decomposition in the form of 1 and 0 that results in secret input of the system as a serial of bits. Prepare the Quranic text including manipulating the text as the UTF file format before embedding to make it compatible with the ASCII code and further control. Open text file to check if there are vowel letters and then save them in vector. Match secret bits with the condition (inserting a kashida in the next letter after the vowel). If matching greater than secret bits then invert the secret bits that are obtained from the watermark ( Apply the embedding process by adding the kashida character when the secret bit is 1 otherwise do not add it. Embed to satisfy the condition of the first vowel letter to hold the secret bit (as kashida). After finishing the cover text, prepare the watermark text to send the receiver.
Embedment in the proposed scheme with the condition is necessary to fulfil the agreement between the sender and receiver as shown in
There are two cases when inserting the bits from the watermark into cover Quranic text which is 1 and 0 depending on the presence or absence of kashida. These four cases control the most embedding issue as summarized in
Secret bit | Kashida within the cover text | Results |
---|---|---|
1 | 1 | 1 |
0 | 1 | 0 |
1 | 0 | 1 |
0 | 0 | 0 |
Results can be interpreted as 1 if the kashida is present in the watermarked text; otherwise, if it is absent, it is represented as 0. The kashida in the original text (cover text) may or may not appear inside. Each byte in the binary image is represented by bit 1 (white) and 0 (black) and these bits are stored in vector to embed it sequentially into the cover text. After locating the vowel letters and insertion into the cover text, the technique checks the condition of the embedding of the letter in the next vowel letter.
Although the extraction process includes many special cases, it is the reverse process of the embedding technique. The enclosed complex data inside the watermarked text needs to be arranged according to the embedding and extracting techniques which may be changed by the external effects. The extracting technique is responsible for the action of all these details. The main contribution of the proposed embedding or concealing technique of the secret bits is the insertion of the kashida at the next letter after a vowel. For extracting the hidden information, one needs to find the vowel letter to locate the next letter. To find the appropriate kashida position for extracting, the condition for the embedding needs to be checked where the presence or absence of the word’s kashida is represented by the secret bit 1 or 0 as shown in
During the transmission, the watermarked Quranic text suffers from many kinds of attacks and tamper detection is one of them. Tamper detection in the proposed scheme aims to find any external manipulation that is unauthorized. The recipient does not know whether the watermarked text was tampered with or not. Thus, the proposed scheme detects whether or not the watermarked Quranic text has been tampered with by reading each letter or word throughout the document. Therefore, that which the watermarked text is converted into consists of the decimal value and vector of the decimal value and is examined by taking the highest value found in the text. Next, by examining and analyzing it with the remaining text characters, one of its values is obtained. A file must be opened to store these values where, afterwards, they are read and combined with different techniques to both stores in a vector and also determine their location as well.
Furthermore, the detection of a tampered-with watermarked Quranic text usually shows that the text has been manipulated by the intruder in one or all of three ways. First, the intruder adds either a single letter to the word or adds a single word to the statement to confuse the recipient of the watermarked Quranic text. Second, the intruder deletes one letter in the word or one word in the sentence, which affects the meaning of the sentence. However, quite often, it is an unimportant letter in the text that the intruder seeks to modify. Moreover, the deletion or shifting of such a letter does not change the whole meaning of the word or statement. Depending on the nature of the language, it is important to work with some letters that have no meaning or are often aesthetically located within the text. Third, the intruder often manipulates the letters in the word or sentence to trick the receiver and becomes successful due to the similar-looking shape and size of the original text in the watermarked Quran. If detected the file must be read again, otherwise, the scheme is claimed to be successful.
In the present study, the proposed scheme finds there are three types of tampering and resolves the drawback of the existing methods for tamper detection inside the program as mentioned above. Thus, two techniques are included in the proposed scheme for controlling tamper detection. These include the queuing technique with the FIFO process and the XOR operation to check changes in the watermarked Quranic text as discussed below.
The queue looks like a 2D array where the rows represent a place of words with each cell consisting of a letter. Every word located in a single row, along with the number of rows in this array, considers the whole text. The last bit in the row specifies the results of the XOR operation, where the last bit in the vector that denotes the XOR of all the bits in the end or rows mentioned before. On the receiver side, the whole text is received as one package and the receiver reads it word for word or even letter for letter. Consequently, the architecture of the text of the Quran looks like queue letter by letter representing the word (raw of the queue) and word by word to represent the statement (verses) or the whole text (column of the queue). The receiver manipulates the text of the Quran as a serial of bits. Therefore, the absence of one letter or word immediately gets discovered due to the change in the sequence of the queue. The receiver side can catch the original text queue by comparing with the key that contains an index for this queue. However, the tampering or changing of the text of the Quran is easy to discover with its secret information.
The secret key that sends to the receiver should contain the indexing of the queue, wherein the indexing queue consists of the manipulated bits generated from many processes based on the XOR operation. Generally, half of the words (50%) are taken from each row and the XOR with another half of the same word is performed to produce the bits considered as the indexing bits. The same procedure is followed for the rest of the queue and applies to all of the text of the Quran as illustrated in
In the queue, the bits of the word XOR results in the bit store at the end of the row which is checked later by the receiver. Changing or modifying any letter of the word leads to an error in this bit. The same XOR operation is applied to all these bits for all the text (queue) and the results are stored in the secret key.
The proposed scheme detects the tampering in the watermarked file by reading each letter or word throughout the document. First, the watermarked text is converted into ASCII and the examined by taking the highest value found in the text. Next, by examining and analyzing it with the remaining text characters one of its values is obtained. A file must be opened to store these values and then they are read as well as combined with the XOR to store in a vector. In this event if the tampered text is detected then the file must be read again; otherwise, the scheme is claimed to be successful. Thus,
Any text file consists of a group of letters and even space is considered a letter under a certain value. Thus, each text file considers a series of letters that can be viewed as a long vector. The letter or character is coded in the computer in a unique ASCII code which is shown as the decimal value. In fact, a text file is a group of decimal value and each value matches one character in the text file. Consequently, for decoding the letters within text file the decimal value file is created similar to the 2D array or matrix with rows and columns as depicted in
The conversion of the watermarked Quranic text file into the corresponding decimal value is necessary. The decimal values are easy to generate using logical operations such as the XOR, OR, AND, and so on. The basic insight can be obtained from the decimal file which is necessary to secure the process and information itself. If conditions are favourable, the logical process is started. In the proposed scheme, this term refers to the selection of the maximum number or value from the decimal file. This maximum number that is used for the important logical operation is called the Exclusive OR process (
Moreover, it is worth explaining the details of the XOR operation here. This operation is considered to be the logical operation that accepts two arguments including the digital input and produces one argument result. For more than two arguments one must combine many operations. The XOR operation is mainly applied due to its ease of use, understanding, and ability to acquire input as a result in the second iteration. It implies that if the output result is XOR-ed again with one input argument then one can get the second input argument. On other hand, if only one of p or q is T/1, then the result will be T/1. Otherwise when both
After the embedding phase, the authentication process starts with the decimal file and the maximum number of some procedure via the XOR operation. Every number in the decimal file is XOR-ed with the maximum number that is derived from the decimal file for tracking. Consequently, the Queue vector is designed to store the results (
After the receiver obtains the necessary information from the watermarked key, the XOR operation is performed to logically generate the queue vector and max value number. This operation results in a decimal file identical to the one derived from the watermarked file. Thus, the procedure of authentication in the system operates like
After finishing the XOR process, the resultant file is compared with the decimal file derived from the watermarked text to identify the presence (or absence) of any differences, indicating the presence (absence) of tampering. This method is very secured, less time consuming and makes it more difficult for the hackers or intruders to guess the secret key. Using this technique, three main threats can effectively be detected such as the insertion of the outlandish character or deletion of one of the specific existing characters, or the modification or reordering of characters in the text file as explained underneath.
In addition,
Every watermarking system aims to evaluate the performance of the resultant watermark media on the sender’s side before sending it to the receiver, thereby ensuring the robustness of the proposed scheme against attacks. In this paper, two types of assessment evaluation are used to check the performance of the proposed scheme: First, the Peak Signal-to-Noise Ratio (PSNR) is evaluated to determine the imperceptibility of the watermarked version. The PSNR value reflects the similarities between the watermarked Quranic text with the original. If the PSNR is high then it is easy to deceive the intruder or attacker, implying the absence of secret bit involvement in this media.
The distribution of the secret bit embedded to cover the Quranic text reflects the PSNR value, while the PSNR detect the frequency of the bits within the cover Quranic text. When the bits become heterogeneous or more chaotic, then the PSNR will get high value. Thus, for embedding is introduced. The high PSNR value indicates good quality of the Quranic text, thereby all the previous methods tried to enhance the PSNR [
With,
Since the MSE double variable adversely affects the double variable PSNR, then there will surely be some inevitable harmful effects. The PSNR parameter allows for the normalizing of the equation for all methods and text types. However, measured either as a ratio or as a(
The second evaluation of the effective performance evaluation of system security for the Quranic text watermarking is the calculation of Normalized Cross-Correlation (NCC) measure. The double variable NCC value reflects the ability of the watermark to withstand against any attack that happened to the watermarked text [
where
The range of the NCC value is in between 0 and 1 and the watermark is said to be the best when it is very close to 1, indicating the closeness of the extracted watermark character variable (
However, the material of data used in the Quranic text dataset was acquired from the public domain via the Internet. The Arabic text was obtained from different reported art-of-the-studies for benchmarking. This dataset was selected due to the enclosure of the sensitive data related to Islam. In addition, surahs of different sizes were chosen for the compression of the host text, benchmarking, and evaluation as explained below.
The implementation of the proposed scheme has been achieved via the use of MATLAB software. In order to evaluate the performance of the proposed approach, it has been considered the standard and authentic version of Quran datasets from
The system is tested by modifying the output file of the watermarked Quranic text to see if the system can detect this case as illustrated in
In
Watermark/byte | ||
---|---|---|
Host size | 1000 bytes | 1000 bytes |
Evaluation scheme | PSNR | NCC |
[ |
52.05 dB | NA |
[ |
62.46 dB | NA |
[ |
NA | 0.98% |
Proposed scheme | 68.20 dB | 1.00% |
Improvement scheme | 9.60% | 1.76% |
The NCC value of the proposed method for the watermark host text was 0.9976%, showing the security enhancement of 1.76% compared to the latest related work 0.98%. In (H) host text, the NCC value for the proposed method was higher than the reported finding which is 0.9976%, for the still (H) host text. Thus, the proposed scheme achieved about 1.76% security and accuracy enhancement for the NCC value compared to that reported in the literature.
The results of the measurement performance using PSNR and NCC in this study to improve the accuracy and security are shown in
The results of NCC as shown in
The benchmark of tamper detection of the proposed scheme compared to other researchers work in the tamper detection attack of the Quranic text-domain is shown in
Reference | Host | Tamper detection | Tamper location |
---|---|---|---|
[ |
Quran | ✓ | NA |
[ |
Arabic | ✓ | NA |
[ |
Quran | ✓ | NA |
Proposed scheme | Quran | ✓ | ✓ |
Tamper detection is a different procedure from other evaluation criteria in terms of the measures where detection gives the condition of true or false. In the simulation program, the notification is received to ascertain the presence or absence of the tamperage. In addition, can be determined the location of the tamper detection inside the word or statement called tamper identification so this process is working when the text has been tampered with.
Indeed, the efficiency of this scheme that we are focusing on the authentication of the tamper detection and location inside the text watermark than existing works. Therefore, we used 3 datasets to get the best experimental analysis results compared with the benchmark of the currently existing work.
The purpose of this paper is to enhance the watermarking scheme of the text of the Quran based on a hybrid technique (XOR and queuing techniques). The methodology consisted of four phases: The first phase was the pre-processing phase followed by the embedding process phase which is performed in order to hide the data after the vowel letters where if the secret bit is “1”, it inserts the kashida but does not insert the kashida if the bit is “0”. The third phase is an extraction process and the last phase is an evaluation of the performance of the proposed scheme by using PSNR for the imperceptibility and NCC for the security of the watermarking. However, in terms of tampering detection, accuracy falls below deletion, insertion, and reordering attacks. The experimental results revealed the improvement of the NCC by 1.76%, PSNR by 9.6% compared to available current schemes. For the future work, researchers should consider the watermarking of the Quranic text when manipulating the text in different formats (.docx, .rtf, and so on) which was out of the scope of this research. Again, some measures that are not mentioned in evaluating the present scheme might also be considered. Many attacks which are not considered in the proposed scheme must be also be looked into. In sum, it is worthwhile to enhance the security and capacity of the scheme for its further betterment.
The authors would like to thank those who contributed to the article and who support them from Universiti Teknologi Malaysia (UTM) for their education.